|Home | About | Journals | Submit | Contact Us | Français|
All strains of Neisseria gonorrhoeae with reduced susceptibility to ceftriaxone and cefixime (cephalosporin-intermediate-resistant [Cephi] strains) contain a mosaic penA allele encoding penicillin-binding protein 2 (PBP 2) with nearly 60 amino acid differences compared to the sequence of wild-type PBP 2, together with a set of resistance determinants (i.e., mtrR, penB, and/or ponA1) that are required for high-level penicillin resistance. To define the individual contributions of these determinants to reduced susceptibility to ceftriaxone and cefixime, we created isogenic strains containing the mosaic penA allele from the Cephi strain 35/02 (penA35) together with one or more of the other resistance determinants and determined the MICs of penicillin G, ceftriaxone, and cefixime. The majority of cefixime resistance is conferred by the penA35 allele, with only a small contribution coming from mtrR and penB, whereas ceftriaxone resistance is nearly equally dependent upon mtrR and penB. Unlike high-level penicillin resistance, the ponA1 allele does not appear to be important for Cephi. A strain containing all four determinants has increased resistance to ceftriaxone and cefixime but not to the levels that the donor Cephi strain does, suggesting that Cephi strains, similar to high-level-penicillin-resistant strains, contain an additional unknown determinant that is required to reach donor levels of resistance. Our data also suggest that the original Cephi strains arose from the transformation of penA genes from commensal Neisseria species into a penicillin-resistant strain already harboring mtrR, penB, ponA1, and the unknown gene(s) involved in high-level penicillin resistance.
Neisseria gonorrhoeae is an obligate human pathogen that is the etiologic agent of the sexually transmitted infection gonorrhea. In 2007, over 350,000 cases of gonorrhea were reported in the United States, but it is estimated that the actual number of infections may be nearly twice that number (5). N. gonorrhoeae primarily infects mucosal surfaces in the lower genital tract, with females often being asymptomatic. If it is left untreated, gonorrhea can cause more serious infections, such as pelvic inflammatory disease and disseminated gonococcal infections, leading to sterility and even death. Since no vaccine is currently available for N. gonorrhoeae, antibiotic therapy is the mainstay for the treatment of infections.
For nearly 40 years, penicillin was the antibiotic of choice for the treatment of gonorrhea, until resistance reached levels that necessitated a switch to other antibiotics. Gonococci utilize two main mechanisms of resistance: plasmid-mediated expression of a TEM-1-like β-lactamase and acquisition of chromosomally mediated resistance determinants that collectively render the organism resistant to clinically achievable levels of penicillin (3). During the same time, tetracycline was also removed as a treatment option, as strains of N. gonorrhoeae with either plasmid-mediated expression of the protective TetM determinant (21) or chromosomally mediated mutations that conferred resistance to the antibiotic (17) had emerged. One of the antibiotics used in lieu of penicillin and tetracycline, ciprofloxacin, was effective for several years, but resistance to this antibiotic also emerged. In 2007, fluoroquinolones were also removed from the recommended list of antibiotics to be used for the treatment of gonococcal infections in the United States (6).
The mechanisms of chromosomally mediated penicillin resistance in the gonococci are complex and multifaceted. Early studies, most notably by Sparling and colleagues (12, 30) and Dougherty (10), demonstrated a stepwise transfer of resistance from a penicillin-resistant strain (e.g., strain FA6140) to a penicillin-susceptible strain (e.g., strain FA19) via DNA uptake and homologous recombination. At least five resistance determinants, which are mutated alleles of normal genes, have been identified and characterized (16, 23, 25, 28, 31). Each of these determinants incrementally increases resistance on its own, but it is the cumulative effects of these determinants and their synergistic interactions that result in clinical levels of resistance. The first step in the transfer of resistance is the acquisition of mutated alleles of the penA gene, which encodes penicillin-binding protein 2 (PBP 2), the primary target of penicillin for lethality. The hallmark of these penA alleles is a single codon insertion, Asp345a, and four to eight mutations clustered in the C terminus of the protein, which together lower the rate of acylation of PBP 2 with penicillin G by 16-fold (26). The second step of transformation results from mutations in the promoter region or the coding sequence of mtrR, a transcriptional repressor that regulates the expression of the MtrC-MtrD-MtrE efflux pump (and other genes ). mtrR mutations result in overexpression of the efflux pump, high-level resistance to hydrophobic agents (such as detergents and dyes), and small (approximately twofold) increases in the MIC of penicillin (16). The third step in resistance involves penB, which encodes mutated alleles of the major outer membrane porin PorB1b with amino acid substitutions in the constriction loop that presumably decrease the influx of the antibiotic into the periplasm (14, 23). Interestingly, the increase in resistance conferred by penB occurs only in strains harboring a coresident mtrR mutation, but the molecular basis for this requirement is not completely understood (24).
Although the aforementioned determinants markedly increase the MIC of penicillin G (from 0.01 to 0.75 μg/ml) when they are transformed into FA19, the MIC is still well below the level for the clinical isolates that serve as DNA donors in transformation experiments (MIC 4 μg/ml). Moreover, DNA from these donors is not capable of further transforming the third-step transformant to higher levels of resistance (3, 10, 28). We demonstrated that high-level-penicillin-resistant strains contain a ponA allele (ponA1) encoding PBP 1 with a missense mutation (L421P) that lowers the rate of acylation with penicillin G by approximately fourfold (28). Surprisingly, whereas the reversion of the sequence of the ponA1 allele back to the sequence of the wild-type ponA gene in high-level-penicillin-resistant clinical strains decreased the MIC of penicillin G twofold, replacement of the wild-type ponA allele with ponA1 in the third-step transformant did not increase the MIC of penicillin G. These data suggest that ponA1 does have a role in penicillin resistance but requires an additional unknown gene(s) to exert its phenotypic effect on resistance. All attempts thus far to identify this gene(s) have been unsuccessful.
Gonococcal infections are now mostly treated with expanded-spectrum cephalosporins, such as ceftriaxone and cefixime. Recently, however, strains of N. gonorrhoeae showing intermediate-level resistance to these antibiotics have emerged, and if (or when) they develop high-level resistance, they will pose a major threat to public health. Thus far, treatment failures have been reported with cefixime but not with ceftriaxone (9, 34). Cephalosporin-intermediate resistance (Cephi) is chromosomally mediated, and the mechanisms resemble those for chromosomally mediated penicillin resistance. The hallmark of Cephi strains is the presence of mosaic penA alleles, which encode PBP 2 variants with 50 to 60 amino acid changes compared to the sequence of the wild-type allele. These mosaic alleles were generated by recombination of regions of the penA genes from N. sicca, N. perflava, N. cinerea, and/or N. flavescens into the gonococcal penA gene (1). Aside from penA, however, the genetics of Cephi have not been comprehensively studied and are mostly unknown. In the present study, we examined the roles of the mtrR, penB, and ponA1 alleles in increasing resistance to ceftriaxone and cefixime. Our data reveal that the mechanisms by which strains have increased resistance to the two cephalosporins are different but that both require the presence of the unknown determinant(s) necessary for high-level penicillin resistance.
The strains and plasmids used in this study are shown in Table Table1.1. FA19 (a penicillin- and cephalosporin-susceptible laboratory strain) and FA6140 (a high-level-penicillin-resistant isolate) were obtained from Fred Sparling, University of North Carolina at Chapel Hill. Strain 35/02 (a penicillin-resistant and Cephi strain) was isolated in Sweden in 2002 (19). pUC18us-penA35 contains the penA gene from strain 35/02, along with 300 bp of downstream sequence and an uptake sequence (11) to facilitate homologous recombination. Transformants harboring the mosaic penA allele were selected with 0.02 μg/ml (strain FA19) or 0.08 μg/ml (strain FA6140) cefixime. The transfer of mtrR from strain 35/02 was accomplished by selecting transformants on 1 mg/ml Triton X-100, and penB35 transformants were selected on 0.37 μg/ml penicillin G; both of these transformations were accomplished with PCR-amplified fragments. The mtrR gene and promoter from strain 35/02 contain both a G45D mutation in the coding sequence of MtrR, which disrupts the DNA-binding domain of the MtrR repressor, and an A deletion (−A) in the 13-bp inverted repeat in the mtrR promoter (19). The −A deletion results in maximal transcription of mtrCDE, whereas the G45D mutation, when it is present by itself, also increases the level of mtrCDE transcription, albeit to a lower level than the promoter mutation (35). pMO-porB1bFA1090, which was used to revert penB back to porB1b, contains a 2-kb fragment of the porB1b gene from strain FA1090 (porB1bFA1090) with an erythromycin resistance cassette downstream of the gene to aid with selection. Plasmids pPR16 and pPR17, which harbor the coding regions of the wild-type and mutant ponA genes, respectively, have an extra 546 bp of downstream sequence to facilitate homologous recombination and the Ω fragment encoding spectinomycin and streptomycin resistance (27) inserted 68 bp downstream of the ponA stop codon to aid with selection (28).
Transformation experiments were carried out essentially as described by Ropp et al. (28). Briefly, piliated colonies of the recipient strain were passaged on a fresh GC medium base (GCB) agar plate and grown for 16 to 20 h. The cells were swabbed from the plate; resuspended in GCB broth with supplements I and II (18), 20 mM bicarbonate, and 10 mM MgCl2; and diluted to an optical density at 560 nm of 0.18. Cells (0.9 ml) were incubated for 5 h at 37°C in a CO2 incubator with 10 μl of a PCR product or plasmid (~3 to 5 μg), and aliquots of the cells were plated on GCB agar plates containing the appropriate antibiotics and allowed to grow overnight. Transformants were passaged on GCB agar plates, and the next day, colonies were boiled in 30 μl of water and spun briefly to pellet the debris. Two microliters of the supernatants was used as the template in a PCR, and transformants showing correct recombination were identified by sequencing.
The MICs of penicillin G, ceftriaxone, and cefixime were determined as described previously (28). Briefly, nonpiliated colonies were streaked on a GCB agar plate and allowed to grow for 18 h. The cells were swabbed into 2 ml of GCB broth with supplements I and II (16), 20 mM sodium bicarbonate, and 10 mM MgCl2 and diluted to an optical density at 560 nm of 0.18. Five microliters of the cell suspension (~50,000 cells) was spotted on a series of plates containing the following ranges of concentrations of antibiotics in approximately 1.5-fold increments: penicillin G, 0.008 to 8 μg/ml; ceftriaxone, 0.0004 to 0.4 μg/ml; and cefixime, 0.0004 to 0.8 μg/ml. The less than twofold increases for most concentrations in the series allowed a more accurate evaluation of the MICs. At least two colonies (and often up to four colonies) from each transformation, verified by PCR amplification and sequencing, were examined for growth on antibiotic plates. At least three independent MIC experiments were carried out, and the MICs reported here represent the averages of all experiments. Error bars in the figures represent the variations of the three determinations.
To examine directly the contribution of the mosaic penA allele to intermediate-level resistance to ceftriaxone and cefixime, we transformed FA19, a β-lactam antibiotic-susceptible strain, and FA6140, a high- level-penicillin-resistant clinical isolate, with the mosaic penA gene (penA35) from the Cephi strain 35/02, which was isolated in Sweden in 2002 (19). Upon recombination of penA35 into FA19, the MIC of penicillin G increased from 0.012 μg/ml to 0.13 μg/ml, a 10-fold increase (Fig. (Fig.1).1). The MIC of FA19 penA35 was twofold higher than the MIC for FA19 transformed with the penA gene (penA4) from FA6140 (MIC = 0.06 μg/ml), which harbors the Asp345a insertion and four C-terminal mutations (24, 26). When the penA35 allele was transformed into FA6140, the MIC of penicillin G increased twofold, from 3 to 6 μg/ml, consistent with the twofold difference in MICs between FA19 penA4 and FA19 penA35. Upon acquisition of penA35, the MIC of ceftriaxone increased 20-fold in FA19 (from 0.0006 to 0.012 μg/ml) and 3.5-fold (from 0.06 to 0.21 μg/ml) in FA6140. The penA35 allele conferred even larger increases in the MIC of cefixime: 100-fold in FA19 (from 0.001 to 0.1 μg/ml) and 10-fold in FA6140 (from 0.048 to 0.5 μg/ml). Importantly, the MICs of ceftriaxone and cefixime for FA19 penA35, although they were markedly increased over those for FA19, were still substantially lower than their respective MICs for 35/02 (especially for ceftriaxone), demonstrating that other resistance determinants play a role in Cephi.
Strain 35/02 has all of the known resistance determinants found in strain FA6140, i.e., penA, mtrR, penB, and ponA1 (19). Although the sequences of the penA genes are markedly different, the sequences of the mtrR and ponA1 determinants in the two strains are 100% identical, and the sequences of the penB genes are over 95% identical, with 100% identity within loop 3, which contains the amino acids most relevant for conferring resistance (23). Thus, we expected that FA6140 penA35 and 35/02, which also have the same penA allele, would have the same level of resistance to the β-lactam antibiotics; however, the MICs of all three antibiotics for FA6140 penA35 were approximately twofold greater than those for 35/02 (Fig. (Fig.1).1). These data suggest either that FA6140 contains an additional determinant(s) not present in 35/02 or that the unknown resistance determinant(s) in FA6140 has a stronger phenotype than that in 35/02.
The synergistic effects of mtrR and penB on penicillin resistance have been well documented (24, 33, 37). When they are present individually in strain FA19 penA4, these determinants have either a small effect (mtrR) or no effect (penB) on the MIC of penicillin G, whereas together they increase the MIC by ~8- to 10-fold. To examine the role of mtrR and penB in Cephi, we carried out a stepwise transfer of the mtrR and penB alleles from strain 35/02 into strain FA19 penA35 (Table (Table1)1) and determined the MICs of penicillin G, ceftriaxone, and cefixime for the resulting transformants (Fig. (Fig.2).2). Consistent with previous results, the transformation of mtrR into FA19 penA35 increased the MIC of penicillin G ~1.5-fold, and further acquisition of penB35 resulted in an additional fourfold increase in resistance. Transfer of the mtrR gene into FA19 penA35 had little to no effect on the MIC of ceftriaxone, but upon the further transfer of penB35, the level of resistance increased about 2.5-fold. In marked contrast to the other antibiotics, neither mtrR nor penB increased the level of cefixime resistance above that for FA19 penA35.
Although overexpression of the MtrC-MtrD-MtrE efflux pump (as a result of the mtrR promoter mutation) has only minimal effects on the penicillin G resistance of strain FA19 penA35 (Fig. (Fig.2),2), the resistance to penicillin G is markedly decreased when the MtrC-MtrD-MtrE efflux pump is inactivated in a high-level-penicillin-resistant clinical strain, indicating that the efflux pump is a critical element of high-level penicillin resistance (33). To investigate further the role of the overexpressed MtrC-MtrD-MtrE efflux pump in Cephi, we knocked out the efflux pump by transforming strains FA6140, FA6140 penA35, and 35/02 with a construct of the mtrD gene (16) disrupted by insertion of the kpt resistance gene (mtrD::Kan) (Table (Table1).1). Inactivation of mtrD in all three strains decreased the MIC of penicillin G nearly 16-fold (Fig. (Fig.3;3; Table Table2),2), highlighting the importance of the MtrC-MtrD-MtrE efflux pump in high-level penicillin resistance. Ceftriaxone intermediate resistance was slightly less dependent on the efflux pump, with an approximately eightfold decrease in the MIC for all three strains being detected when the pump was disrupted. In contrast, the disruption of mtrD conferred a much smaller decrease (two- to threefold) in the level of cefixime resistance, which was consistent with the apparent lack of an effect of either mtrR or penB on cefixime resistance, as described above (Fig. (Fig.22).
In contrast to the role of mtrR, the role of penB in high-level penicillin resistance is less clear. The synergism between mtrR and penB suggests that at least part of the role of mtrR, in addition to increasing the efflux of antibiotics, is to promote the resistance phenotype of penB, which is silent in the absence of mtrR (24). To distinguish between its role in efflux and the activation of PenB, we reverted the penB gene of FA6140 penA35 back to the wild-type porB1bFA1090 gene by allelic exchange and determined the MICs of penicillin G, ceftriaxone, and cefixime for the resulting transformants. Reversion of the penB allele back to the wild-type conferred consistent two- to threefold decreases in the MICs of each of the three antibiotics (Fig. (Fig.33 and Table Table2).2). For both penicillin G and ceftriaxone, reversion of penB to wild-type porB1b did not decrease the MIC as much as deletion of the MtrC-MtrD-MtrE efflux pump, whereas for cefixime, both reversion to wild-type porB1b and deletion of the efflux pump had similar effects on the MICs.
We also investigated the role of ponA1, which encodes PBP 1 containing a single missense mutation (L421P), in Cephi either by transforming a third-level transformant with ponA1 or by reverting the sequence of the ponA1 allele in FA6140 penA35 and 35/02 back to the wild-type sequence. As previously reported for penicillin resistance, transformation of ponA1 into a third-level transformant had no appreciable effect on resistance to any of the three antibiotics tested (Fig. (Fig.2).2). When the sequence of the ponA1 allele was reverted back to the wild-type sequence in FA6140 penA35 and 35/02, the level of penicillin G resistance decreased twofold, consistent with our previous data (Fig. (Fig.4)4) (28). However, reversion of the ponA allele had little to no effect on ceftriaxone or cefixime resistance. Thus, in contrast to high-level resistance to penicillin, ponA1 appears to have little or no role in Cephi (Fig. (Fig.44).
The data presented in this report provide insight into the individual contributions of the mosaic penA35 allele and the mtrR, penB, and ponA1 resistance determinants to Cephi. We showed that mosaic penA alleles (e.g., penA35) play an important role in Cephi but that the other determinants, with the exception of ponA1, are also required. A useful way of examining the individual contributions of the penA35 allele versus the contributions of the entire set of other determinants is to compare the MICs of ceftriaxone and cefixime for three strains: FA19, FA19 penA35, and FA6140 penA35 (Table (Table3).3). FA19 is an antibiotic-susceptible strain that serves as the starting point for transformation studies, FA19 penA35 shows the increase in resistance due solely to the penA35 allele, and FA6140 penA35 shows the sum of the effects of penA35 in the presence of the complete set of determinants critical for high-level penicillin resistance.
The MICs of both antibiotics increased ~400-fold from FA19 to FA6140 penA35 (Table (Table3),3), but how this 400-fold increase was achieved was markedly different for the two antibiotics. For ceftriaxone, transfer of the penA35 allele into FA19 conferred a 20-fold increase in the MIC, while the additional determinants in FA6140 caused another 20-fold increase in MIC, for a total increase of 400-fold. Thus, the mosaic penA allele and the other resistance determinants contribute equally to the increase in the MIC for this antibiotic. In contrast, transfer of penA35 into FA19 increased the MIC of cefixime by 100-fold, while the other determinants in FA6140 caused only an additional 4-fold increase in the MIC. Thus, while the overall increases in the MICs of the two antibiotics are the same, it is clear that for cefixime, the penA35 allele is the major factor in resistance, while for ceftriaxone, the additional determinants, which function at least in part by limiting the concentration of the antibiotic in the periplasm, are as important as the mosaic penA allele in increasing the MIC. These data demonstrate the different modes by which these two cephalosporins inhibit the growth of gonococci.
Our data can be compared to those reported by Takahata et al. (32), who transformed FA1090 with mosaic penA constructs and determined the MICs of a range of β-lactam antibiotics. In that study, the MICs of ceftriaxone and cefixime were increased 4- and 15-fold, respectively, upon transfer of the entire mosaic penA allele into FA1090. Those increases are significantly lower than the increases reported here in FA19 (20- and 100-fold). The differences in the two studies in part may reflect the fact that FA1090 contains a penA gene (NGO1542 of the sequence with GenBank accession no. AE004969) with an Asp345a codon insertion that decreases the rate of acylation of PBP 2 with β-lactams (29), which would increase the MIC for the starting strain and thus lower the increase in the MIC for the resulting mosaic penA transformant.
Ochiai et al. (22) determined the 50% inhibitory concentrations of a range of antibiotics for inhibition of Bocillin FL (a fluorescent β-lactam) binding to purified PBP 2 variants from two antibiotic-resistant strains, one with a penA allele similar to that of strain FA6140 and the other with a penA allele similar to that of strain 35/02. They reported that the 50% inhibitory concentrations of ceftriaxone for the two PBP 2 variants were not different, inferring that the mosaic penA allele plays little to no role in ceftriaxone intermediate resistance. In contrast, our data show that isogenic replacement of the nonmosaic penA allele in FA6140 with the mosaic penA35 allele increases the MIC of ceftriaxone fourfold, from 0.05 to 0.2 μg/ml (Fig. (Fig.1).1). These results clearly indicate that the mosaic penA35 confers an increase in the ceftriaxone MIC over and above that conferred by penA4. The reasons for this discrepancy are not clear but may be due to the utilization of inhibition curves to define inhibition instead of the derivation of k2/Ks acylation constants for the various antibiotics. Although in vitro binding data are useful for comparison of the effects of mutations on PBPs, we believe that the analysis of isogenic strains created with resistance alleles whose sequences are known is the most valid and straightforward approach to examination of the contributions of specific alleles in conferring resistance to antibiotics.
Comparison of the MICs of the three antibiotics for FA19 penA35 with or without mtrR and penB mutations reveals that the contributions of these two determinants to resistance vary depending upon the antibiotic (Fig. (Fig.2).2). Thus, transfer of both mtrR and penB into FA19 penA35 increased the MIC of penicillin G 6-fold and that of ceftriaxone 3.3-fold but had little to no effect on the MIC of cefixime (Fig. (Fig.1).1). The latter result was unexpected, as these determinants would be predicted to decrease the level of permeation of any antibiotic that enters the periplasm through porin channels. Thus, these data suggest either that cefixime does not diffuse into the periplasm through porin channels or that such diffusion is not altered by penB mutations. One obvious distinction of cefixime from penicillin G and ceftriaxone is that cefixime has a −2 net charge at neutral pH, whereas penicillin G and ceftriaxone have a −1 net charge. The−2 net charge may affect the permeation of cefixime.
Our data also shed light on the dual roles of the mtrR determinant on Cephi and highlight the different modes of action of the three antibiotics. mtrR mutations affect resistance in two ways: increased efflux of the antibiotic and activation of penB mutations that decrease diffusion through porin channels. In contrast, penB mutations increase reistance only when an mtrR mutation is present. Deletion of mtrD in FA6140 penA35 or 35/02 caused marked 16- and 8-fold decreases in the MICs of penicillin G and ceftriaxone, respectively, whereas the MIC of cefixime was less affected (three-fold decrease) (Fig. (Fig.33 and Table Table2).2). Conversely, the reversion of the penB sequence back to the wild-type porB1b sequence decreased the MICs of all three antibiotics about the same: between two- and threefold. These data suggest that the increases in resistance to penicillin and ceftriaxone upon the acquisition of the mtrR and penB determinants are due to a combination of efflux by the MtrC-MtrD-MtrE pump and a decrease in permeation through porins. In contrast, the mtrR mutation appears to contribute to cefixime resistance solely by activating penB, since the decreases in MICs following the inactivation of mtrD and the reversion of the penB sequence to the wild-type sequence were highly similar. The fact that these genetic manipulations had any effect at all was surprising, since we were unable to observe a change in the MIC of cefixime when mtrR and penB were transferred into FA19 penA35 (Fig. (Fig.3).3). The reason for this discrepancy is not known, but it may reflect the influence of the unknown determinant in FA6140 on both the efflux pump and PenB.
Unlike previous studies, which associated penA35, mtrR, penB, and ponA1 with Cephi through epidemiological and sequencing studies (19, 36), this study directly establishes the contributions of these resistance determinants in Cephi. However, these determinants were not sufficient to increase the MICs of any of the three antibiotics to donor levels when they were transformed into strain FA19. Attempts to transform the level of resistance of FA19 penA35 mtrR penB ponA1 to donor levels with 35/02 or FA6140 genomic DNA were unsuccessful (data not shown). These data on penicillin resistance are identical to those previously observed by us and others and suggest the presence of an unknown determinant(s) that is nontransferrable by transformation and that modulates increased levels of resistance (10, 28).
The role of ponA1 in high-level penicillin resistance is well established, as shown by the twofold decrease in the MIC upon the reversion of the ponA1 sequence to the ponA wild-type sequence (Fig. (Fig.4)4) (28). However, no such decrease in the MIC was observed for ceftriaxone or cefixime, indicating that ponA1 plays little to no role in resistance to these two antibiotics. This distinction may reflect the differences in the rates of acylation of the two essential PBPs, PBP 1 and PBP 2, by the three antibiotics. For the wild-type PBPs, PBP 2 has an ~16-fold higher acylation rate with penicillin G than PBP 1 (2, 28), making PBP 2 the primary lethal target for this antibiotic in wild-type strains. Upon acquisition of the penA allele, the rate of acylation of PBP 2 decreases by ~16-fold, bringing it much closer to that of PBP 1 (although PBP 2 is still the lethal target, since the introduction of an altered PBP 1 with a lower rate of acylation for the antibiotic does not increase the MIC of penicillin). Upon acquisition of the other resistance determinants, the lethal target of penicillin must shift to include PBP 1, and thus, acquisition of an altered PBP 1 increases the MIC. For the cephalosporins, we suspect that there may be much larger differences in their rates of acylation of PBP 2 compared to those of PBP 1, such that even with the decrease in the rate of acylation of mosaic PBP 2 with the cephalosporins and the presence of the other resistance determinants, PBP 1 never becomes a lethal target, and accordingly, alterations in PBP 1 have no effect on resistance. Experiments to examine the rates of acylation of the various wild-type and mutant PBPs with the three antibiotics to test this possibility are in progress.
In conclusion, the present study clearly emphasizes that the mosaic penA allele, mtrR, penB, and an unknown mutation, but not ponA1, are important determinants for conferring intermediate resistance to the expanded-spectrum cephalosporins ceftriaxone and cefixime. Together, these determinants confer 400-fold increases in the MICs of both antibiotics but do so in different ways: the mosaic penA allele and the set of other resistance determinants contribute equally to resistance to ceftriaxone, while the mosaic penA allele confers the greatest increase in resistance to cefixime. The inability of ponA1 to confer resistance to strain 35/02 or strain FA6140 penA35, even though it is found in a wide range of Cephi strains (19), suggests that these strains did not arise de novo from antibiotic-sensitive strains but likely arose by transfer of mosaic penA alleles into preexisting chromosomally mediated penicillin-resistant strains, which persist in the bacterial population, even though penicillin has not been recommended for use for the treatment of gonorrhea for nearly 25 years (4). This would explain the presence of the ponA1 allele, which is required for high-level penicillin resistance, in almost all Cephi strains, even though it does not contribute to Cephi. Such a mechanism would also account for the rapid emergence of these strains over the last several years. The current MIC breakpoint for strains considered resistant to either ceftriaxone or cefixime is >0.25 μg/ml. Given the documented ability of N. gonorrhoeae to acquire additional mechanisms of resistance to β-lactam antibiotics, it appears to be a matter of when and not if strains resistant to these antibiotics will emerge, and when they do, the dearth of other antibiotics useful for the treatment of gonococcal infections portends a looming public health problem.
This work was supported by grants AI36901 (to R.A.N.) and GM66861 (to C.D.) from the National Institutes of Health and a grant from the Research Committee of Örebro County, the Örebro University Hospital Research Foundation, Örebro, Sweden (to M.U.).
Published ahead of print on 15 June 2009.